INDEX

Explanations

Q and A starters

The neuron fires on standalone decimal numbers (floating‐point numerals) embedded in the text.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 this

-2.67

 when

-1.98

 that

-1.91

on

-1.81

for

-1.78

 because

-1.62

 first

-1.59

 decorativas

-1.56

 these

-1.56

 When

-1.53

POSITIVE LOGITS

1.91

"*

1.75

//"

1.73

an

1.73

↵↵↵↵↵↵↵↵↵↵↵↵

1.55

,"

1.55

possibly

1.55

{"

1.55

AND

1.52

":

1.50

Activations Density 0.013%